ML17256A778
| ML17256A778 | |
| Person / Time | |
|---|---|
| Site: | Ginna  |
| Issue date: | 06/09/1983 |
| From: | Chen C, Fulton J, Herr J GILBERT/COMMONWEALTH, INC. (FORMERLY GILBERT ASSOCIAT |
| To: | |
| Shared Package | |
| ML17256A777 | List: |
| References | |
| NUDOCS 8306200316 | |
| Download: ML17256A778 (67) | |
Text
I'r'o June 9, 1983 ROBERT E.
GINNA NUCLEAR POWER STATION CONTAINMENT VESSEL TENDONS
RESPONSE
TO USNRC REVIEW COMMENTS ON TENDON'VALUATION PREPARED FOR ROCHESTER GAS AND ELECTRIC COMPANY WRITTEN BY:
F. Fulton REVIEWED BY:
C.
He r APPROVED BY:
C.
en 830b2003ib 830b13 PDR ADOCK 05000244' PDRj PREPARED BY:
GILBERT/COMMONWEALTH READING~ PENNSYLVANIA Qbert ICoemoeuealth
~glN MNEf HLfl',OPl
TABLE OF CONTENTS.
SECTION TITLE PAGE
1.0 INTRODUCTION
2.0 ROCK ANCHORS 2.1 2.1.1 2.1.2 2.1.3 2.1.4 SMALL SCALE TESTS Rock Anchor C
Rock Anchor A Rock Anchor B
Conclusions 2.2 2.2.1 2.2.2 2.2.3 FULL-SCALE ROCK ANCHORS Rock Anchors 46 64 58 and 61 Sam le Rock Anchors Conclusions on Full-Scale Rock Anchors 9
10 14 15 3.0
RESPONSE
TO REMAINING NRC COMMENTS 18 REFERENCES TABLE 1 Bond Lengths Resulting From Stressing Actual Rock Anchors FIGURES 1-25 Gdbert /Commonwealth
C INTRODUCTION The report entitled "Robert E. Ginna Nuclear Power Station, Containment Building Tendon Investigation" (Reference
- 1) describes an investigation into the possible causes for the lower-than-predicted tendon forces which were measured during. past surveillances on the Ginna containment.
This report has been.
reviewed by staff members of the Operating Reactors Branch of the U.S. Nuclear Regulatory Commission (2),
and the enclosure to Reference 2 summarizes the results of the NRC staff review.
The staff concluded that stress relaxation of the tendon wires as the only single cause for the larger-than-predicted force loss was not conclusively demonstrated in the tendon report (1).
Six items of concern are identified which form the basis for this conclusion.
Items 1, 2, and 3 pertain to the tests on the three small scale rock anchors reported in the FSAR and to the four full size rock anchors selected for initial testing.
The remaining items pertain to:
the comparison of the elongations of the Ginna wall tendons with those experienced by wall tendons in a concrete containment not employing rock anchors (item 4); the stress relaxation test results (item 5); and the results from the 1981 tendon surveillance (item 6).
The main concern of the staff deals with the integrity of the rock anchors.
Therefore, to address this concern the results of the tests on the rock anchors and the results measured during their original tensioning are evaluated in further detail in Section 2.0.
This discussion also specifically addresses the staff concerns in items 1, 2 and 3.
The three remaining items are responded to on an item-by-item basis in Section 3.0.
Additional comments by the staff, items 7 through 12 in Reference 2, are also addressed in Section 3.0.
Qbert /Commonwealth
2.0 ROCK ANCHORS 2.1 SMALL SCALE TESTS In April of 1966, tests were conducted on three small scale rock anchors (A, B, and C) ~
Each anchor consisted of twenty-eight; 1/4-inch diameter wires.
The lengths of anchors A, B, and C were 12'-8 3/4", 13'-2 1/2" and 14'-8 3/4", respectively.
The specified grouted length was 4'-5 1/2" for anchors A and C, and 5'-5" for anchor B.
All three anchors were installed in 3 1/2 inch diameter -holes.
The results of these tests are shown in FSAR Figures 5.6.1-6 thru 5.6.1-8, and the tests are discussed in the accompanying FSAR section.
This information was summarized in section 2.1.1 of Reference 1.
The test results for each of these small scale anchors are evaluated in further detail below.
2.1.1 Rock Anchor C (Figure 1)
This test anchor was used to demonstrate the design bond stresses for the actual 90-wire rock anchors.
For the actual anchors, a
design grout length of 22'-0" was established so that at the maximum tensioning force of approximately 850 kips (corresponding to 80/ GUTS) the shear stress at the grout-rock interface averaged over the grouted length would not exceed 170 psi.
In the test, the tensioning force was applied in 20 kip increments, and dial guages were used to measure the movement of the anchorhead.
The recorded data points are marked on the results shown in FSAR Figure 5.6.1-8, which is included as Figure 1 herein.
Also shown in Figure 1 are two predicted load-elongation lines.
The line designated as L = 10'-3 1/4" corresponds to the ungrouted anchor length, obtained from the total anchor length of 14'-8 3/4" minus the specified grout length of 4'-5 1/2".
Data points lying on this line would indicate that the anchor wi res remained firmly bonded to grout and that no slip
&!bertICommonwealth
occurred.
The line designated as L = 14'-8 3/4" corresponds to the total anchor length.
Data points which lie on this line would indicate that the bond between the anchor wires and grout was completely broken and that the entire tensioning force had been transferred to the lower anchorhead, which remained fixed.
From Figure 1, at 20 kips applied to the anchor, the test results indicate that the anchor wires remained bonded to the grout.
At some value of force between 20 kips and 40 kips, the bond between the wires and grout became completely broken over the entire grouted length, and up to 100 kips the applied force was entirely resisted by the lower anchorhead which remained fixed.
Beyond 100 kips, the elongation of the anchor exceeded that based on its total length, which indicated that the lower anchorhead experienced a gradual vertical displacement as the load was increased.
This displacement probably occured as a result of bearing stresses which were high enough to cause the anchorhead to crush the grout.
The anchor continued to resist the applied force without slip until 200 kips was reached, when slip first occurred.
Just before the slip, the lower anchorhead displacement was 0.22 inches as noted in Figure 1.
Beyond 200 kips, and up to the maximum force applied of 264 kips (80% GUTS), the anchor continued to resist the applied force with accompanying slips.
Some key bond stresses are noted on the measured results.
The stresses abw and obr are the calculated average bond stresses over the grouted length of 4'-5 1/2" at the wire-grout interface and at the rock-grout interface, respectively.
Therefore,'etween abw of 17 psi and 34 psi the bond between the wires and grout became comPletely broken.
At an average bond stress abr of 170 Psi between the rock and grout, movement of the lower anchorhead began at 100 kips.
Even with this movement,
- however, the load-elongation relationship is nearly linear out to 200 kips.
At this load the average bond stress was 340 psi at the rock-grout interface.
The test anchor continued to develop the force beyond
&tbert/Commonwealth
200 kips up to the maximum specified force of 264 kips, where the corresponding stress at the grout-rock interface was 448 psi.
However, at each 20 kip plateau beyond 200 kips, the recorded movement of the upper anchorhead indicated a slip of the anchor under constant load.
It was concluded originally from the test results that the actual rock anchors, designed for 170 psi shear stress at the grout-rock interface, would have a factor of safety of 2.0 against anchor slippage.
Additionally, from the discussion above, it is seen that:
(1) bond between the wires and grout was completely broken early in the loading process when abw was between 17 psi and 34 psi; (2) the test anchor continued to resist the applied force up to a bond stress at the grout-rock interface of 170 psi with no movement of the lower anchorhead, thus indicating that neither crushing of the grout nor slip at the grout-rock interface occurred; and (3) beyond 170 psi some gradual movement of the lower anchorhead
- occurred, but the anchor was able to develop the applied force without slippage until the bond stress at the grout-rock interface reached 340 psi and anchor slip started to occur.
The behavior of this test anchor will be compared with the response of the other two small scale anchors
- and, subsequently, with load-elongation results for a 10X sample of the 160 rock anchors.
For this 10X sample, the bond stresses occurring at the maximum stressing force of 80/
GUTS are noted on the test results in Figure 1.
At this force, the average bond stress over a grout depth of 23 feet (average of 16 anchors) was calculated to be 44 psi along the wire-grout interface and 163 psi along the grout-rock interface.
Gdbert /Commonwealth
2.1.2 Rock Anchor A (Figure 2)
This anchor was selected for two tests, designated as A-1 and A-2.
Prior to 100 kips Then the set up.
the start of test A-l, the anchor was tensioned to in one load increment and shimmed (point 1 in Figure 2).
lifting frame shown in FSAR Figure 5.6.1-5 (Figure 3) was The anchor was loaded in 20 kip increments until visual lift-offwas recorded at 110 kips.
The measured results are shown by the closely spaced solid and dashed lines in Figure 2.
At 110 kips, the design hold-down capacity of the rock assumed in the design of the actual rock anchors was demonstrated, as discussed in the FSAR.
This test was designated as Test A-1.
The load was then transferred from the lifting frame to the concrete pier under the anchor for Test A-2.
In Test A-2, the jacking force was increased in 20 kip increments until 80%
GUTS (264 kips) was reached.
The recorded data points are indicated in Figure 2.
As discussed in the FSAR, the dashed line joining the first data point in Test A-2 with point 1 occurs due to the transfer of the jack reaction from the jacking frame, used for Test A-l, to the concrete pier under the anchor.
This apparent discontinuity would not be evident had the test anchor been loaded continuously from zero force to 264 kips with the jack reacting against the pier throughout the test.
Therefore, to evaluate the test results, it is necessary to shift the curve marked "Test A-2" so that the elongation measured at 100 kips is consisted with that measured at 100 kips as the test anchor was initially stressed.
From the curve noted as "Test A-2 Shifted",
several conclusions can be drawn as to the bond development mechanism in this anchor.
~First, two predicted load-displacement curves are superimposed in the test results in Figure 2.
The line designated as L = 8'-3 1/4" is the response of the anchor based on its unbonded length protruding from the grout.
This assumes that there was Gilbert/Commonwealth
perfect bond and no slip occurred between the wires and the grout.
The second liney L 12 8 3/4 y assumes that the bond between the wires and grout was completely broken and, consequently, the force applied to the anchor was transferred directly to the fixed lower anchorhead.
The fact that data point 1 lies on the L = 12'-8 3/4" line indicates that by the time the force reached 100 kips, all the bond between the wires and grout had been broken.
Since data was not required to be recorded between 0 and 100 kips, from this test alone it would be impossible to determine the value of force at which the bond breakage actually started.
However, since the.
specified grouted length of this anchor, 4'-5 1/2", was the same as for test anchor C, it is reasonable to expect that the bond between the wires and the grout was probably broken somewhere between 17 psi and 34 psi.
Beyond 100 kips, the results are similar to those measured in test C.
The measured displacement.of the upper anchorhead exceeded that predicted, indicating that movement of the lower anchorhead was occurring.
Also, as in test C, first slip of the anchor was noted at 200 kips, when the lower anchorhead had displaced 0.23" (see Figure 2).
- However, slippage of the anchor at and above this force level was much smaller for test A-2 than for test C.
In general, the performance of the anchor in test A-2 was very consistent with that in test C,
and the factor of 2 against anchor slippage was confirmed.
2.1.3 Rock Anchor B (Figure 4)
The measured displacements of the upper anchorhead exceeded the predicted values throughout the test.
At the first 20 kip force increment, the measured displacement exceeded that predicted even
<or the assumption that the anchor wires were completely unbonded from the grout (L = 13'-2 1/2").
From a comparison with the L =
13'-2 1/2" line in Figure 4, the lower anchorhead appears to have experienced an initial displacement of approximately 1/8-inch at GilbertICommonwealth
20 kips; however, the results seem to indicate that progressive movement of the lower anchorhead did not begin before 100 kips.
This is evidenced by the approximately uniform 1/8-inch offset of the measured displacement between 20 kips and 100 kips.
Beyond 100 kips, the measured displacements exceeded the predicted values by greater amounts at each load increment, which indicates that grout crushing or perhaps some slippage at the grout-rock interface was progressively occurring.
The first apparent slip of the entire anchor occurred at 180 kips, and only an additional 20 kips could be applied before the test was halted because of excessive slip.
Although not pointed out in the FSAR discussion of the test
- results, there appears to be an explanation as to why this anchor behaved so differently from anchors A and C.
A search through project correspondence revealed that test anchor B was observed to actually have a grouted length of only two feet when removed, rather than the S'-S" specified.
As a result of this, grouting procedures were established for the actual full-scale rock anchors in order to prevent a similar occurence;
- and, based on the measured response of the actual anchors during tensioning, it appears that these procedures were entirely effective.
The results on the full-scale anchors are evaluated in Section 2.2.
As far as the test on anchor B is concerned, the results are not considered to be representative.
However, it is interesting to note that even with the insufficient grout length of test anchor B, this anchor was able to develop a significant force (180 kips) before progressive slippage of the entire anchor was recorded.
Conclusions In the three small scale tests, the short grouted length of the wires was incapable of developing even small forces in bond, and the applied force was transferred almost immediately to the lower Qbert /Commonwealth
anchorhead.
For test anchor B, this condition appears to have occurred almost at the onset of loading and was accompanied by movement of the lower anchorhead due to an insufficient grout length.
The actual grout length is believed to have been approximately two feet instead of the 5'-5" length specified.
Because of this, the load-displacement response of the anchor was more non-linear throughout the entire loading range, although apparent slip of the entire anchor did not occur until the applied force exceeded 180 kips.
The behavior of this anchor is attributed to its insufficient grout length and the results are considered to be unrepresentative.
Nevertheless, the test was useful in that it prompted a review of the grouting procedure for the full-scale anchors.
Test anchors A and C exhibited a much more predictable behavior than anchor B.
These anchors had grouted lengths of 4'-5 1/2",
which were also too short a distance for the applied load to be developed in bond between the wire and the grout.
- However, once the load was transferred to the lo~er anchorhead, both anchors responded linearly up to 100 kips.
Beyond this point, which corresponds to an average bond stress between the grout and rock of 170 psi, movement of the lower anchorhead occurred in a gradual manner until first slip of the anchors occurred at 200 kips or 340 psi bond stress between the grout and rock.
This implies that the full-scale anchors, which were designed for 170 psi at their maximum stressing
- force, had a safety factor of 2 against anchor slippage.
In summary, for all three small scale test anchors the relatively short grouted length of the anchor wires caused the stressing force on the anchor to be transferred to the lower anchorhead.
- However, the fact that the grout lengths of the test anchors A and C were about twice that of anchor B, resulted in a more linear response of these anchors.
A safety factor of 2 against anchor slippage of the full-scale rock anchors was determined from the tests on A and C.
Qbett /Commonwealth
Considering the evaluation of the small scale tests described
- above, the comments of NRC staff members in items 1 and 2 of Reference 2 have been addressed.
From the evaluation it is seen that the load level at which initial slip occurred was consistent with the grouted lengths of these anchors.
Anchors A and C, which both had a grout length of 4'-5 1/2", experienced slip at the same load level, 200 kips.
Anchor B apparently had a grout length'f only two feet, and it slipped at a lower load, 180 kips.
In
.addition, the load-deformation relationships were very similar for anchors A and C, while that for anchor B was considerably different due to its reduced grout length.
The degree of linearity of the load-deformation curves was explained by loss of bond at the wire-grout interface and by movement of the lower anchorhead, both of which occurred at lower load levels for anchor B than for anchors A and C.
2.2 FULL-SCALE ROCK ANCHORS During the stressing of each of the 160 rock anchors, the position of the stressing ram was recorded at approximately 1000 psig increments on the pressure gauge attached to the ram.
Each increment corresponded to approximately 130 kips force on the anchor.
From this data, the elongation response of each anchor can be determined.
In addition, four of these rock anchors, numbers 46, 61, 58, and 64, were designated for more detailed elongation measurements using dial gauges to record the movement of the upper anchorhead.
The results from these four anchors, and from a lOX sample of all 160 anchors, were discussed in section 2.1.2 of Reference 1.
The results are examined in more detail in the following discussion.
Gdbert ICommoowealth
2.2.1 Rock Anchors 46 64 58 and 61 Rock Anchor 46 (Figure 5)
This anchor, as well as all others, is 33'-10 1/2" long and consists of 90 1/4-inch diameter wires.
The anchor had a free length of 9'-9" (second stage grouted after the anchor was stressed) and a first stage grouted length of 24'-1 1/2".
After the first stage grout had developed its minimum specified compressive
- strength, rock anchor 46 was stressed up to 743 kips (70% GUTS) and unloaded.
Then the anchor was stressed to 849 kips (80% GUTS),
shims were inserted, and the anchor was seated at 731 kips.
Dial gauges recorded the movement of the upper anchorhead during these loading cycles and the results are shown in Figure 5.
Also shown in the figure are the predicted force-elongation curves of the anchor.
The line L =9'-9" corresponds to the length of the anchor that protrudes from the grout.
The line L = 33'-10 1/2" is the predicted response of the anchor assuming that the bond between the wires and the grout was completely broken and that the lower anchorhead remained fixed in resisting the entire force applied to the anchor.
For this anchor, the force in the Initial Tensioning Test was applied in 100 kip increments.
The measured elongations (circled dots) were greater than those predicted, assuming a condition of complete bond and no slip between the wires and grout over the entire grouted length (L = 9'-9" line).
- However, the measured elongations were much less than those corresponding 'to the L = 33'-10 1/2" line.
This indicates that, unlike the small scale
- anchors, the wires remained bonded to the grout along some distance above the bottom anchorhead and the applied force was not transferred to the lower'nchorhead.
The elongations tend to follow the dashed line denoted as LEq = 15'-8".
This line represents the equivalent free length of the anchor wires, including the 9'-9" actual free length.
Therefore, the difference Qbett I Commonwealth 10
in the anchor length of 33'-10 1/2" and LEq = 15'-8", or 18'-2 1/2", represents a measure of the length of the anchor adjacent to the lower anchorhead that remained bonded to the grout.
The corresponding length of anchor in the grout which would be completely unbonded is equal to the equivalent length of 15'-8" minus the 9'-9" free length, or 5'-ll".
These distances are denoted in Figure 6, where in this case LCUB = 5'-ll",
LCB1 = 18'-2 1/2", LEq = 15'-8",
LC = 24'-1 1/2", and LF = 9'-9".
The 18'-2 1/2" distance is referred to as a measure of the bonded length because actually over the 5'-ll" portion of the equivalent free length embedded in the grout, the wires are not expected to be
~corn letel unbonded from the grout.
Slippage could have occured at this interface, but shear stresses of some magnitude were probably developed and sustained.
Consequently, the relative slippage between the wires and grout was not as large as would occur if the wires were completely free from the grout.
Because of this, to get the same total displacement of the upper anchorhead corresponding to the 15'-8" equivalent free length, slippage between the wires and the grout in the presence of shear stresses at this interface must occur over an embedded length of the anchor which is greater than 5'-ll".
This increased length is referred herein to as the partially unbonded length which is denoted by LpUB in Figure 6.
- Here, one possible idealization of the corresponding bond stress distribution at the wire-grout interface is also indicated.
Due to the uncertainty in the actual distribution of bond stresses for the multi-element rock anchor, an accurate determination of LpUB cannot be made.
- However, from the results in Figure 5, it is clear that a length of anchor remaining completely bonded to the grout (LCB2 in Figure 6) does exist since the measured elongations are considerably less than those based on the total anchor length.
A rough estimate of LpUB and LCB2 can be made, assuming the idealized bond stress distribution in Figure 6. If the idealized Qbeft ICommonwealth 11
curve is approximated by a uniform distribution, LpUB and the average bond stress are uniquely determined from the force applied to the anchor and the corresponding displacement of the upper anchorhead.
This calculation for rock anchor 46, based on its force at 80%
GUTS of 849 kips and 1.234 inches measured displacement, gives an average bond stress of 83 psi occuring-over a length LpUE = 12.05 ft.
This results in a length LG82 = 12.08 ft. over which the wires remained completely bonded to the grout and where no bond shear stresses existed along the wire-grout interface.
In effect, these results indicate that the anchor was long enough that a force of 849 kips could be developed without relying on the lower anchorhead.
In addition to the estimate on bond length, a practical criteria for determining the acceptance of the measured elongations of rock anchors in general is recommended in Reference 3.
From section 4.3.8 of this reference, the displacement of the rock anchor measured during stressing is considered to be acceptable if the equivalent free length (LEq) does not exceed the sum of the actual free length (LF) and 50% of the grouted length (LG).
In other words, LEq must not extend more than halfway down into the grouted length of the anchor.
Applying this criteria results in a value of 21.82 ft. for LF + 0.5 LG, and since LEq = 15.67 ft., the acceptance criteria was met for this anchor.
The results presented above have been tabulated in Table 1 for comparison.
Similar results are provided in the table for the remaining anchors to be discussed.
After the anchor was unloaded (triangle dots in Figure 5), it was stressed to 80%
GUTS (849 kips) and then anchored at 731 kips.
The elongation of the anchor during the Final Stressing is indicated by the solid dots in Figure 5. It is evident that the anchor displacements closely followed the first loading cycle.
G~tbett /Commonwealth 12
This indicates that there was no degradation of the anchor, nor was there any apparent increase in the unbonded or partially bonded lengths along the anchor wires from that observed in the Initial Tensioning Test.
These results are considered as a
significant indication that the rock anchor performed very well during these loading operations.
Rock Anchors 64 58 61 (Figures 7, 8, and 9)
Each of these anchors was stressed to a force level near 80%
GUTS (849 kips), and the displacement of the upper anchorhead was measured using a dial gauge at specified pressure levels along the way.
Shims were then inserted and the anchor was seated.
The results due to stressing are shown in Figure 7 thru 9.
The results are similar to those obtained for anchor 46 in that the measured elongations lie close to the LEq line and are greater than those based on the free length of the anchor alone but less than the elongations predicted based on the total length of the anchor.
- Thus, bond breakage and slip between the wires and the grout along some length of the anchor had occurred, but the anchor remained completely bonded and unstressed over some length above the lower anchorhead.
Estimates of the lengths of the extent of bond are included in Table 1.
Column 6 in the table shows the values for LEq range from 18.00 ft to 21.25 ft.
These are less than the values corresponding to the acceptance limits of LF + 0.5 LG, which range from 21.65 ft. to 22.98 ft.
Finally, the length of each anchor estimated to ranging from remain completely bonded above the lower anchorhead, 12.08 ft. to 3.52 ft., is shown in Column 11. It is important to point out that since the anchors contain lower anchorheads, they are expected to be capable of developing the applied force even if the bond between the wires and the grout were completely broken over the entire length of the anchor.
This was demonstrated in the tests on small scale anchors A and C.
Gilbert /Commonwealth 13
2.2.2 Sam le Rock Anchors (Figures 10 through 25)
Every tenth rock anchor spaced around the circumference of the containment was selected for review as discussed in section 2.1.2 of Reference 1.
The data recorded during stressing of this 10%
sample is examined below.
By way of a clarification, the holes for the anchors were numbered consecutively around the circumference, but the numbers assigned to the rock anchors themselves were different from their hole designations (see Table 1).
The wall tendons were numbered the same as the holes.
During the stressing of the rock anchors, the position of the stress'ing ram was recorded initially at 500 psi gauge pressure and then at 1000 psi pressure increments up to 80% GUTS, which corresponds to 6560 psi gauge pressure or 849 kips.
The data is plotted in Figures 10 through 25.
The ram position at the 500 psi and 1500 psi pressure level is extrapolated back to determine the ram position at zero pressure, and this serves as the origin of the two "predicted" lines corresponding to the free length of the anchor (LF) and the total anchor length (LT).
For either the actual data or the predicted response lines, the difference in ram position at a given pressure and the position at zero pressure (for example 0.65 inches for anchor 68 in Figure 10) represents the elongation of the anchor.
The results shown in the figures are similar to those observed for the full size anchors discussed previously.
The measured elongations trend in a linear manner, and no erratic behavior of any of the anchors is indicated.
The non-linear behavior observed for the small scale test anchors at their higher load levels does not'occur for the actual anchors.
The small scale tests on A and C showed that by the time the load level was reached which produced an average bond stress between the grout and the rock of 170 psi, complete bond breakage between the wires and the grout had already occurred and movement of the lower anchorhead had Gilbert ICommonweaIth started.
However for the full-scale anchors, at 170 psi average bond stress between the grout and the rock (which occurs at 80%%d GUTS), the bond between the wires and the grout was not completely broken as evidenced by the fact that the elongations of the rock anchors were still considerably less than those based on the total anchor length.
The equivalent free lengths of the anchors (LEq) range between 18.00 ft. and 19.92 ft., as shown in Column 6 of Table l.
All values are less than the values corresponding to the acceptance limit of LF + 0.5 LG, which range from 21.85 ft. to 22.60 ft.
The estimated length along each anchor where the bond between the wire and the grout was partially broken (LFU8) or remained unbroken (LCB2) is indicated in Columns 10 and ll, respectively.
The average bond stress calculated over LpU8 appears in Column 12.
For the 10% sample, the estimated anchor lengths that remained completely bonded above the lower anchorhead range from 4.67 ft.
to 8.72 ft.
- However, as pointed out previously, the rock anchors do not require reliance on bond between the wires and the grout since they are equipped with bottom anchorheads.
2.2.3 Conclusions on Full-Scale Rock Anchors The force-elongation data recorded during the stressing of the rock anchors indicates that these anchors performed very well.
One anchor,
- 846, was subjected to two loading cycles and the force-elongation results were quite reproducible.
The anchors responded in a nearly linear manner under the applied stressing force.
The slight non-linearity exhibited in most of the curves indicates bond slip between the anchor wires and the grout, which is normal.
Slip of the full-scale anchors in their respective holes did not occur, as it did for the small scale anchors.
Due to the bond slip between the wires and the grout, the measured elongations were greater than those corresponding to the free Gtbert
/Commonwealth 15
length of the anchors protruding from the grout.
- However, the actual elongations were considerably less than those which would have occurred had bond breakage progressed all the way to the lower anchorhead.
This indicates that complete bond between the wires and the grout existed for some length above the lower anchorhead for each rock anchor.
Calculational estimates of the bond lengths confirm that this condition existed, although complete bond is not required for the rock anchors to effectively develop the stressing force since each anchor is equipped with a bottom anchorhead.
Based on the actual elongations, the effective free length was calculated for each anchor and found, in every case, to satisfy accepted acceptance criteria which requires that the effective free length not extend more than halfway down into the grouted length of the anchor.
Referring to the NRC staff comments in items 2 and 3 of Reference 2, the information presented above. shows that the load-deformation relationships were considerably more linear for the full-scale anchors than for the small scale
- anchors, particularly anchor B.
The full-scale anchors resisted the 0.80 GUTS maximum stressing force by bond at the wire-grout interface.
However, for the small scale anchors, this bond was completely broken and the force was transferred to the lower anchorhead.
Subsequently, the anchorhead displaced, and finally a force was applied which was large enough to cause the grout plug to slip in the rock.
This sequence of events did not occur for the full-scale anchors; consequently, their load-deformation relationships were much more linear, as shown.
The non-linearities in the load deformation curves of the small scale anchors have been explained in the above evaluation, and they did not occur as a result of time effects, e.g.,
creep of the grout or rock.
Qbett /Commonwealth Relative to the NRC comment in item 3, as a result of this further investigation, it is now clear that the difference in measured and predicted elongations for the full-scale anchors were due to a partial loss of bond at the wire-grout interface rather than slip at the grout-rock interface, as stated on page 2-3 of the report (1).
The amount by which the measured elongations exceeded the values predicted, based on the free anchor length alone, are within acceptance criteria established for rock 'anchors in Reference 3.
h Cnthert /Commonwealth 17-
3 '
RESPONSE
TO REMAINING NRC COMMENTS (4)
Staff Comment:
The comparison of measured to predicted elongations for "rock anchored tendons" and "non-rock anchored tendons" (Table 2.1-1) indicates that the prediction of the elongations for "rock anchored tendons" is much more uncertain than for "non-rock anchored tendons."
g~es onse:
The results in the table shows the opposite to be the case.
The results indicate that during the original stressing of the wall tendons at Ginna in April 1969, the measured elongations were equal to their predicted values for 16.9X of the tendons, which is greater than the corresponding percentage of 9.6X for the non-rock anchored tendons.
Therefore, the uncertainty was less, not more, for Ginna.
- However, the primary purpose of the table is not to demonstrate that the Ginna tendon elongations were more predictable than those for the non-rock anchored sysem.
The ASME Containment Code has for a long time recognized that measured tendon elongations are likely to be somewhat more or less than predicted.
This is reflected in the Code provisions by an acceptance tolerance on measured elongations of plus or minus 10/
of those predicted.
The evaluation in Table 2.1-1 was performed to determine if there is any evidence which would indicate that. during the stressing of the wall tendons at Ginna, measurable slip of the rock anchors might have occurred.
If such slip had occurred, then it would Qbgft ICOmmanwgalth 18
have raised a question as to whether the slip could possibly have continued after the wall tendons were stressed a condition would have provided an explanation
/.
lower-than-predicted tendon forces measured at and anchored.
Such of the subsequent surveillances.
Slip of the rock anchors would have resulted in movement of their upper anchorheads, which are coupled to the bottom anchorheads of the wall tendons.
This condition would have evidenced itself in the elongations of the wall tendons being greater than predicted for an unusually la'rge percentage of the
- tendons, compared with the frequency at which it normally occurs.
In addition for these
- tendons, the amount by which measured elongations would have exceeded those predicted would have been much greater than that which normally occurs.
The norm used for this comparison is a non-rock anchored tendon system.
The data in Table 2.1-1 shows that the frequency of measured elongations exceeding predicted values was significantly less for Ginna than for the non-rock anchored system.
For Ginna, only 40.6X of the wall tendons had actual elongations in excess of their predicted values.
This is one-half of the 80.0X of the non-anchored wall tendons that were in this category.
From these results, it is clear that during the original stressing of the Ginna tendons, the frequency at which measured elongations exceeded predicted was far less than would be expected had rock anchor slip occurred.
In addition, as seen from Table 2.1-2, the measured elongations of the Ginna tendons~
which did exceed their predicted values did so by percentages which are entirely acceptable.
These values are well within the 10K Code allowable; but, more importantly, the values are not significantly different from those of the non-rock anchored tendons.
An explanation as to why rock anchor slip would not be expected to have occurred during wall tendon stressing is provided in section 2.1.2 of the report.
As discussed, the stressing of the wall tendon does not change the stress in the rock anchor until Qbert /Commonwealth 19
all of the force on the upper anchorhead of the rock anchor is transferred from its bearing plate to the wall tendon anchorhead.
As indicated in the report, a conservatively low value of 0.60 GUTS was assumed for force in the rock anchor.
Until this force level is reached in the wall tendon, there is no increase in the force in the rock anchor to even potentially cause slip.
Thus, in stressing the wall tendons to 0.80
- GUTS, 75X (corresponding to 0.60 GUTS) of this force is applied without producing any change in the force in the rock anchor.
The remaining 25% of wall tendon force is developed primarily in the second stage grout, which had not been stressed prior to this time.
The second stage grout length, being in the neighborhood of 10 ft., is sufficient to resist the force increment from 0.60 GUTS to 0.80 GUTS in the wall tendons without increasing the stresses in the first grout stage that resulted from the initial tensioning of the rock anchors.
Therefore, with the wall tendon at 0.80 GUTS, the stresses produced in the second stage grout were low enough so that any slip that might possibly have occurred in the rock anchor was not large enough to have"produced measurable displacement of its anchorhead to which the wall tendon is attached.
Because of this, the elongations of the wall tendons that were measured at Ginna were as small, and at least as predictable, as those normally measured for tendon systems which do not contain rock anchors.
The same conclusions are reached for the 137 tendons at Ginna that were retensioned in June 1980.
As the results in Table 1 of Appendix A in the report
- show, the measured elongations were in good agreement with their predicted values, especially considering the fact that the elongations were small, being about one-third of those being measured during the original stressing of the wall tendons.
In addition, the summary of the data in Table 2.1-1 shows that compared to the non-rock anchor
- tendons, a larger percentge of Ginna tendons had measured elongations that were the same as predicted (14.6X) and a smaller percentage of measured Gdbert ICenmonwealth 20
elongations that were in excess of those predicted (65.0%).
Again, at the June 1980 retensioning of the wall tendons, as at their original stressing in April 1969, there was nothing in the results to indicate that slip of the rock anchors had occurred.
(5)
Staff Comment:
The stress relaxation tests conducted at Lehigh University (Section 3) indicate that two of the three sample wires exhibit greater stress relaxation properties than what was used in the original design while one sample wire exhibits less relaxation.
This does not totally support the licensee's contention that stress relaxation is the sole cause of prestress loss.
~Res ense:
The sample wire referred to as exhibiting stress relaxation less than the design value is from tendon 150 and the test results are shown in Figure 3-2 of the reports As seen from Figure 3-2, the stress relaxation test results are below the design curve (12% wire) for the specimens tested at 0.70 GUTS and at 68 F,
78 F, and 104 F.
For the remaining specimen, Ill which was tested at 0.75 GUTS and 104 F, the test results lie far above the design curve.
As discussed on pages 3-9 and 3-10 of the report, the results for the specimens tested at 68 F and 78 F appear to confirm one another.
Therefore, it seems reasonable to conclude that for these temperatures, the sample wire from this tendon does indeed exhibit a stress relaxation property less than the design curve.
However, the results for the remaining specimen, 813 (curve 150-B) which was tested at
.70 GUTS and 104 F, appear to be disproportionately low considering the much larger stress relaxation results for specimen fkll (curve 150-A).
In light of this, it was concluded that the test results for specimen 813 do not appear to be Qbert IComeonwealS reasonable.
Consequently, as explained on pages 3-12 of the
- report, a stress relaxation curve for the 0.70 GUTS/104 F test condition was estimated for use in the comparison with the measured effective stress relaxation presented in section 2.9.3 of the report.
The estimated curve appears in Figure 2.9-4 of the report.
As discussed on page 2-38 of the report, the actual tendons are expected to have experienced temperatures somewhere in the 85 F to 95 F range for most of their existence in the containment.
If this is considered in evaluating the test results for the wire specimens from tendon 150, the relaxation property of actual tendons comprised of wires from the same heat as these specimens would be expected to exhibit a stress relaxation property greater than the des'ign curve.
(6)
Staff Comment:
The 1981 tendon surveillance results (Table 5-2) show that the ratio of measured to predicted tension loss of same type tendons (from same heat) varied in a wide range, from 1.0 to 3.2.
The 1981 prediction of the tendon force loss should have used the Lehigh test results.
R~es ense:
It appears that the designation "Tendon Type" in Table 5-1 has been misinterpreted.
This does not indicate wire heat.
As stated in section 5.2 of the report "tendon type" designates the shape (curvature) of a tendon.
For example, type A tendons are straight tendons.
At the other extreme, type GS tendons have the largest cummulative curvatures along their length.
This designation scheme was adopted in the original design to categorize the different friction factors for the tendons.
Regarding the staff comment in the last sentence, the effective stress relaxation of each tendon was considered to be the best Qbert ICommoaWRRIth 22
estimate of the relaxation property of a specific tendon.
Many of the tendons in the 1981 surveillance were not comprised of wires from the same heats as the Lehigh test wires.
(7)
Staff Comment:
The proposed future surveillance, Section 7.0, is not adequate in light of the Lehigh test results which indicate that two-thirds of the tendons may lose the design safety margin in about 10,000 hours0 days <br />0 hours <br />0 weeks <br />0 months <br /> due to stress relaxation alone.
Therefore, a more frequent inspection, on an annual basis, is required for this reason as well as that noted above.
R~es ense:
It is not clear why it would be concluded from the Lehigh test results that the tendons could lose their design margin after 10,000 hours0 days <br />0 hours <br />0 weeks <br />0 months <br />.
Nevertheless, this proved not to be the case as the results in the report of the July 1981 surveillance indicated.
This surveillance was conducted at nearly 10,000 hours0 days <br />0 hours <br />0 weeks <br />0 months <br /> after the June 1980 retensioning',
and the results in Table 5-2 show that the force in every tendon was considerably above the 636 kip design requirement, with the average tendon force being 714 kips or 12.3%%d above the design requirement.
Therefore, at this point in time there is nothing to indicate that the surveillance schedule in Section 7.0 will not be adequate.
(8)
Staff Comment:
The licensee should provide the NRC staff with a schedule which indicates when the Lehigh tests for tendon wire relaxation will be available
- and, based on that schedule, when the licensee will provide an assessment of relaxation predictions by the factor method.
The schedule should be provided immediately.
That assessment should be included in the next tendon surveillance Gilbert /Commonwealth report and integrated with the data collected during the next surveillance.
It is expected that a significant portion of the Lehigh test data will be available before the report of the summer 1983 tendon surveillance is required to be forwarded to the staff.
Therefore, there should be sufficient data available in order to construct curves of tendon force vs. time, including tolerance, bands of allowable upper and lower bounds for the tendons surveyed as outlined in Reg.
Guide 1.35.1.
It is expected that the next forthcoming surveillance report will contain such curves so that an extrapolation can finally be made of the tendon forces to be expected in the future.
It is assumed that the next tendon surveillance will be conducted on tendons which include at least some of the same group as those surveyed in the 1981 inspection.
~Res onse:
The Lehigh tests have been completed and the evaluation of the results is in progress.
The retensioned wire results are being evaluated by both the factor method and the, superposition method.
The method which provides the best prediction of the retensioned stress relaxation properties will be used in the prediction of the tendon forces for the summer 1983 surveillance.
These predicted forces will be available in time for the surveillance, tentatively scheduled for mid-July.
Some of the tendons selected for this surveillance will be the same as those surveyed at the July 1981 inspection.
The results of the evaluation of the Lehigh retensioning tests will be provided as part of the report on the 1983 tendon surveillance, which would be issued within 90 days after the completion of the surveillance.
(9)
Staff Comment:
In the discussion of the effects of the neoprene pads on the tendon prestress loss, it appears that this unique feature of the prestressed concrete system was rather summarily dismissed as a
Gd bert /Commonwealth contributing factor to tendon losses.
Specifically, a visual inspection of the pads should be performed even if some removal of concrete is involved.
The thickness of the pad material should be measured and samples obtained for evaluation of the condition of the elastomeric material.
The assumptions made in the report concerning pad deformation were extrapolated from handbook values and must therefore be considered optimistic especially since the pads are overstressed for such material.
Since each 0.1 inch of inelastic deformation in the pads would reduce the tendon force by about 9 KIPS (at the 742 KIP design load), the contribution of the pads to loss of tendon force cannot be readily dismissed.
- Also, since these pads are designed to function as a hinge mechanism in the event of an accident that causes containment pressurization, specific information of their condition should be obtained rather than assuming that they are in serviceable condition.
R~es ense:
The evaluation in the report on the elastic and creep deformation of the neoprene pads is considered to be reasonably complete based on the available test data.
The staff has commented that the calculated creep deformations of the pad "were extropolated from handbook values and must therefore be considered optimistic especially since the pads are overstressed for such material."
The handbook referred to is Reference 4 in the report, and creep curves from that reference were provided as Figure 2.8-2 in the report.
This data is considered to be reliable since it is stated in Reference 4 that the creep curves represent the results of ten-year creep tests on neoprene.
Also, the pads are not overstressed in terms of the design limitation stated in Reference 4.
This requirement is that the initial pad deflection under the bearing stress shall not exceed 15X of the thickness of the pad; and based on the discussion in section 2.8.1 of the
- report, the pads did meet this requirement.
Qbert /Commonwealth 25
(10) Staff Comment:
The report does not reference the Load history which may have been collected by monitoring the four load cells which are installed in the tendon system.
These load cells were used to record tendon force losses collected in original containment leak rate test (GAI Report No.
1720 dated October 3, 1969) and subsequently to continuously monitor tendon forces.
The licensee should provide available data from the load cell readings taken since the retensioning of the tendons and since the 1981 surveillance along with a critical assessment of the information.
R~es ense:
Subsequent to the installation of the four (4) load cells used in the original containment leak rate test, a short-term monitoring program was conducted between March and July, 1981 to assure levels of tendon prestress prior to the scheduled 1981 surveillance.
Another monitoring program.commenced in Augus 1981 and continued until July, 1982.
No monitoring was conducted subsequently.
The results of the two monitoring programs will be reported as part of the report on the 1983 tendon surveillance.
(11) Staff Comment:
The licensee should provide predictions about the expected behavior of the retensioned wire for the following potential conditions.'a)
If the tendons should ever need ta be retensioned again in order to maintain prestress.
(b)
Expected behavior of the retensioned wire in the event of containment pressurization.
Such predicted behavior should be confirmed, to the extent possible, during the next integrated leak rate test.
&Iberia /ComeonWealth 26
R~es onse:
A decision as to whether or not another tendon retensioning will be required cannot be made until the results of the upcoming 1983 surveillance have been evaluated.
The behavior (elongation) of the tendons under a containment pressurization would be an elastic response to the pressure load and is, therefore, not influenced by the fact that the tendons have been retensioned.
The elongation would be expected to be small.
e (12) Staff Comment:
The long-term effects of variation of temperature on the stress relaxation properties of the wire and the tendon system have not been sufficiently addressed.
~Res onse:
The effect of temperature variations on the stress relaxation property was investigated to a far greater extent than is normally done to -establish such p'roperties for tendon wires. It is felt that, considering the data from all the test conditions, the tests were sufficient to prove that the stress relaxation properties of the tendons were greater than the design value.
QtR.rc ICommanWealth 27
I f
n '~
REFERENCES 1.
Robert E. Ginna Nuclear Power Station Containment Buildin Tendon 2.
Letter from D. M. Crutchfield, USNRC/ORB, to J.
E. Maier, RGKE, on Containment Vessel Tendon Evaluation Program, dated March 8, 1983.
Post-Tensionin Manual, Post-Tensioning Institute, Third Edition, 1981
'bert
/Coemonwealth
TABLE 1.
BOND LENGTHS RESULTING FROM STRESSING ACTUAL ROCK ANCHORS ROCK ANCHOR 8 ACTUAL LENGTHS LT LF LG LEq HOLE 8 (ft.)
(ft.)
(ft.)
(ft.)
(2)
(3)
(4)
(5)
(6)
ESTIMATED E UIVALENT LENGTH BOND DISTRIBUTION LCUB LCB1 LF+0.5LG LPUB LCB2 AVE STRESS (ft.)
(ft.)
(ft.)
(ft.)
(ft.)
( si)
(7)
(8)
(9)
(10)
(11)
(12) 46 64 58 61 10X SAMPLE 18 33.88 9.75 24.13 15.67 10 33.88 9.2 24.46 18.00 92 33.88 12.08 21.80 21.25 113 33.88 12.08 21.80 20.92 5.92 18.21 21.82 12.05 12.08 8.58 15.88 21.65 17.10 7.36 9.17 12.63 22.98 18.28 3.52 8.84 12.96 22.98 17.66 4.14 83 59 55 55 68 102 96 31 33.88 11'7 22.71 19.00 7.83 14.88 22.52 15.64 7.07 11 33.88 9.83 24.05 18.08 8.25 15.80 21.85 16.52 7.53 21 33.88 11.33 22'5 19.17 7.84 14.71 22.60 15'70 6.85 61 64 64 17 41 33.88 11.00 22.88 18.58 7.58 15.30 22.44 15.17 7.71 65 29 51 33.88 10.75 23.13 19.42 8.67 14.46 22.32 17.29 5.84 57 1A 21A 61 33.88 10.92 22.96 18.00 7.08 15.88 22.40 14.24 71 33.88 11.00 22.88 18.58 7.58 15.30 22.44 15.17 8.72 7.71 70 65 7A 11A 80A 86 155 90A 81 33.88 10.92 22.96 18.25 7.33 15.63 22.40 14.62 8.34 91 33.88 10.67 23.21 19.42 8.75 14.46 22.27 17.45 5.76 101 33.88 11.00 22.88 19.42 8.42 14.46 22.44 16.85 6.03 111 33.88 11.08 22.80 19.58 8.50 14.30 22.48 16.68 6.12 121 33.88 10.92 22.96 18.08 7.16 15.80 22.40 14.30 8.66 131 33.88 10.75 23.13 19.92 9.17 13.96 22.32 18.38 4.75 68 57 59 59 69 54 139 37A 141 33.88 10.75 23.13 18.00 151 33.88 10.92 22.96 18.92 7.25 15.88 22.32 14.57 8.56 8.00 14.96 22.40 16.05 6.91 68 62 59 160 33.88 10.50 23.38 19.83 9.33 14.05 22.19 18.71 4.67 53
4 CONTAINMENT VESSEL ROCK ANCHOR TEST C
Orv 4
o
~
II O
II Q7 O
e O II Vl U
W O C/l O
W I ~
CL CY CCI W
I O
C4 I-Ln CL')
W W
~ ~ I- )
C O
OOO ~
~
O ~ I-X CY I
CD W A $
CeJ I
I CCI CJ' II I
e CD II C?.'
O CLW O 4 W 0 I
IV CC W
I IJ O
I W
e 0
OIL lU CI ee) a II Ie I~
II Pe eeP rg0 n
ee eee VI ~
~ OV0",
I' e
C'e FIGURE j.
?
CI CONTAINMENT Il'i>SEL ROCK ANCHOR TEST SKEET
)
0
~0 C+
LLI
<<C I
I (X
CQ O N
~ ~
D hC O
O LLI I
Pl M II CV II CV O
~
Pl P)
O CO N O O
D D O Ill D
Irl D
4J CC Ul I
I-Irl UJ UJ I )O l
I LIl I
I(0 I ~
II N
I r
O II I
LL U
O LU l0 D O D
4J IO UJ I
I 0
CC0 Q
W0 vl WxVX rP ~ P U
I I
Z X
W W 0 W
W 0
0 W
'X 0
I W
0 Al I
<<'l ns n
O o
0 C
CI Ct'I O
Q
'O Ct
~0 I
n dl)I OV O'I FIGURE 2
I rp
(
~
~
f-lyomi tc Jwc<
puu -Ro>
LCP WIMc
'L ~
'J I
~)
CQ C)
I I
Ql I>
'ig I slI)
Ill I ~
)AEA~UP't&<Ct.~-
D>'LTUIV7
.=-..;
-A I
~
I 4 g.'-'.
~cd I >(
~4' I
V///y~j/'/ ) ) I g
~ ~ a r(y
~ CIHOgg I
n~
/~XX/ ~y J~
Rec.w ~cue~
Comcwe.~K l~j/r p,o c-x Aec+~ y~sr P'I WR.
.=.p/)-.=//i
CONTAINMENT Ir'ESSEL ROCK ANCHOR TEST
>>4
>><)
- r<
0I pl a>>
U Ul II
~4 a
'N V
O C/I rrI W
II O
O C) p O I
Co W
O Cal I
V W
cQ X O
c 5C O
U ~
I W
IY I
A Ca>>i C>>J
~
I II Cal CO II O Z I
Ii CD
- i. u rv O o
- j. 0 IJ 0 X 0
- l. 0 2.0
- j
- d. b5
- 2. o
- 2. o Zio
- 3. 0
- I
- l f